Scrutineer, in development
What if code written by AI had to pass an independent inspection?
Judges agent-built code, and wrote none of it.
InspectionSample
ChangeNew sign-in flow
TestsReal, not hollow
ScopeStayed within the ask
VerdictOne issue to fix
Scrutineer is in development. Screens show sample content, and details may change.
It judges. It never builds.
Scrutineer wrote none of the code it inspects, so it has nothing to defend.
It judges. It never builds.Sample
Written byAn agent
Inspected byScrutineer
Passed byYou
A verdict you can act on.
Get a clear answer on what's wrong and where, before the change goes anywhere.
A verdict you can act on.Sample
Sign-in redirectFails on expired links
Test coveragePasses
Unrelated editsNone
What we're building in.
Sees only the work
It checks the change against what was asked.
Catches hollow tests
Tests that prove nothing get called out.
Fits before you merge
Inspection comes before the change lands.
What if code written by AI had to pass an independent inspection?
We'll tell you when Scrutineer is ready.